DM-12798: Add source selection by signal-to-noise ratio #95
Conversation
The various source selection modules have had common code to apply minimum and maximum limits. Factor this common code into a base class.
|
|
||
|
|
||
| class SignalToNoiseLimit(BaseLimit): | ||
| """Select sources using a magnitude limit |
|
|
||
| def apply(self, catalog): | ||
| """Apply the magnitude limit to a catalog | ||
|
|
| ---------- | ||
| catalog : `lsst.afw.table.SourceCatalog` | ||
| Source to which to apply limits. | ||
|
|
There was a problem hiding this comment.
Wording: it’s a catalog of sources to which the limit will be/is/gets applied
| Returns | ||
| ------- | ||
| selected : `numpy.ndarray` | ||
| Boolean array indicating whether each source is selected |
There was a problem hiding this comment.
The “each” doesn’t seem right here. Replace with “a given” or “...indicating for each source whether it is selected”?
Also note that sources for which the general flag for the fluxField is set will be excluded? Also, since you allow for the error flux field to be different, should you select against its flux flag as well (if different from fluxField...different flux fields failure modes may differ).
There was a problem hiding this comment.
The flags can be handled by the RequireFlags selection, so no need to include them here.
| maximum = pexConfig.Field(dtype=float, optional=True, doc="Select objects with value less than this") | ||
|
|
||
| def apply(self, values): | ||
| """Apply the limit to an array of values |
|
|
||
| def setDefaults(self): | ||
| pexConfig.Config.setDefaults(self) | ||
| self.flags.bad = ["base_PixelFlags_flag_edge", "base_PixelFlags_flag_interpolated", | ||
| "base_PixelFlags_flag_saturated"] | ||
| self.signalToNoise.fluxField = "base_PsfFlux_flux" |
There was a problem hiding this comment.
Do you also need to default the errField (since having them associated is not assumed?)
| high.set("other_fluxSigma", 0.001) | ||
| self.config.doSignalToNoise = True | ||
| self.config.signalToNoise.fluxField = "other_flux" | ||
| self.config.signalToNoise.errField = "other_fluxSigma" |
There was a problem hiding this comment.
I think @parejkoj is trying to homogenize the use of err vs. Sigma in the stack (RFC-333).
There was a problem hiding this comment.
Sadly, none of that has been implemented yet, so fluxSigma will remain.
Bit better wording, with thanks to Lauren MacArthur.
Masayuki Tanaka (NAOJ) suggested that a signal-to-noise ratio limit would be a useful addition.
PaulPrice
force-pushed
the
tickets/DM-12798
branch
from
c295f05
to
7441574
Compare
If you've got only a magnitude (not flux), you can't use the SignalToNoiseLimit without some modification. However, for magnitudes the error is the inverse of the signal-to-noise ratio, so we can just put a limit on that.
File generated by pytest.
PaulPrice
force-pushed
the
tickets/DM-12798
branch
from
7441574
to
b111fa0
Compare
No description provided.